IBM was the first company to develop "speech technology" that transcribed the spoken word into text. The company’s eight research labs around the world, including the one in India, work in tandem to develop nifty new technologies that aim to make life a lot easier for people, from the BPO industry to the US military.
The company’s India Research Lab, based at the Indian Institute of Technology in Delhi, has developed the "Real Time Language Technology", which has completed its pilot stage. And to find out what the technology was all about, I spoke to Ashish Verma, a senior researcher at the India Research Lab, who spearheaded the project. I began by asking Dr Verma, an eight-year veteran of IBM’s research lab, on the applications of the technology.
"The technology has been developed forpeople for whom English is not the first language, and therefore have a problem with pronunciation and tone of spoken English. Essentially, all Indian languages are flat, but the English language is all about inflection and tone. The technology is interactive and corrects the speaker’s pronunciation, tone, inflection, and grammar," Dr Verma said. "We see this application finding wide usage in the BPO and education sectors," he said.
The efficiency of the technology is being assessed by a human assessor. The IBM Real Time Language Technology, he said, is a Web-enabled interactive language technology for people who speak English as a "second language".
It is based on IBM’s speech recognition and other advanced speech processing techniques, and aims to help people develop their spoken language skills, largely accent-free. Dr Verma said that the technology corrected the syntax, grammar and inflection of the spoken word, enabling the speaker to come up to speed in the correct usage of the English language.
The "Real Time Language Technology" will first be deployed at the IBM Daksh facility in Gurgaon. Apart from correcting the pronunciation and tone, the technology also has voice-enabled grammar evaluation tests. Dr Verma said the tests identify areas for improvement by highlighting shortcomings and providing examples of correct pronunciation and grammar.
The language technology is based on the Windows platform, though it can be equally effective on the Linux open-source systems, he said. Earlier this month, IBM delivered a bi-directional English to Arabic translation software to U.S. forces in Iraq. The software, called Mastor (Multilingual Automatic Speech-to-Speech Translator) aims at improving communication between military personnel and Iraqi forces and citizens.
The company’s India Research Lab, based at the Indian Institute of Technology in Delhi, has developed the "Real Time Language Technology", which has completed its pilot stage. And to find out what the technology was all about, I spoke to Ashish Verma, a senior researcher at the India Research Lab, who spearheaded the project. I began by asking Dr Verma, an eight-year veteran of IBM’s research lab, on the applications of the technology.
"The technology has been developed forpeople for whom English is not the first language, and therefore have a problem with pronunciation and tone of spoken English. Essentially, all Indian languages are flat, but the English language is all about inflection and tone. The technology is interactive and corrects the speaker’s pronunciation, tone, inflection, and grammar," Dr Verma said. "We see this application finding wide usage in the BPO and education sectors," he said.
The efficiency of the technology is being assessed by a human assessor. The IBM Real Time Language Technology, he said, is a Web-enabled interactive language technology for people who speak English as a "second language".
It is based on IBM’s speech recognition and other advanced speech processing techniques, and aims to help people develop their spoken language skills, largely accent-free. Dr Verma said that the technology corrected the syntax, grammar and inflection of the spoken word, enabling the speaker to come up to speed in the correct usage of the English language.
The "Real Time Language Technology" will first be deployed at the IBM Daksh facility in Gurgaon. Apart from correcting the pronunciation and tone, the technology also has voice-enabled grammar evaluation tests. Dr Verma said the tests identify areas for improvement by highlighting shortcomings and providing examples of correct pronunciation and grammar.
The language technology is based on the Windows platform, though it can be equally effective on the Linux open-source systems, he said. Earlier this month, IBM delivered a bi-directional English to Arabic translation software to U.S. forces in Iraq. The software, called Mastor (Multilingual Automatic Speech-to-Speech Translator) aims at improving communication between military personnel and Iraqi forces and citizens.